OBJECTIVE: Discrete choice experiments (DCEs) were introduced to health economics in the 1990s. The last systematic literature review considered papers published before 2017, since then the number of studies and breadth of topics has rapidly grown. Previous reviews identified new directions and methodological challenges, and the last review appealed for better study reporting. Given numerous guidelines and checklists, a review of current state of practice is timely.
METHODS: A systematic literature review of health related DCE studies published between 2018 and 2024. The search strategy and data extraction replicated previous reviews and we report comparable statistics to show how the method has evolved. We also extract new data to show current state of practice for issues not included in previous reviews including area of application, reporting of qualitative data, and details on survey administration, design, and analysis.
RESULTS: Data extraction is on-going and will be complete before the ISPOR conference. Of the 8293 papers identified, 1560 met the inclusion criteria and data are being extracted. The range of research questions addressed, and the geographical settings have expanded, nevertheless eliciting patient preferences is the most common application followed by general population preferences. Despite best practice guidelines, compared to previous reviews fewer studies used qualitative methods to select attributes and levels, and details of these methods are still underreported. Most studies use a D-efficient design, however most of these do not report details on the priors used and utility function specification. Most studies investigate preference heterogeneity linked to respondent characteristics. The use of advanced econometric methods to account for unobserved preference heterogeneity has increased. However, details on the model specifications are poorly reported.
CONCLUSION: The number of papers continues the near exponential growth identified in previous reviews. Despite calls for improved study reporting, many study still report incomplete details for quality assessment.
